Summary On 18 March 2002, the LakeCarling loaded a cargo of iron ore pellets at berth No.2, Sept-les, Quebec, and departed the same day bound for Point Lisas, Trinidad. The next morning during scheduled rounds it was discovered that No.4 hold was taking on water. Further inspection revealed that a six-metre fracture had developed on the port side shell. Sea ice thwarted attempts to keep a collision mat in place to stem water ingress and the bilge pumps were unable to keep up. Additional pumps were brought on board from a Canadian Coast Guard vessel tasked to the area and these were sufficient to stabilize the situation. On 21March2002, the salvage tug RyanLeet arrived on the scene. With the help of more powerful pumps and with the fracture partially plugged from the exterior, No.4 hold was pumped dry. The vessel made its way to the protected waters of the Baie de Gasp where more caulking work was done in way of the fracture. On 26March2002, the vessel weighed anchor for Qubec, Quebec, for permanent repairs. Ce rapport est galement disponible en franais. Other Factual Information Description of the Vessel The LakeCarling is a conventional Handy-sized bulk carrier with bridge, accommodations and engine room located aft of the five cargo holds. The vessel is of the gear-less type, i.e. without its own cargo handling equipment. The main engine drives a single right handed propeller. History of the Voyage On 14 March 2002, the LakeCarling arrived and anchored at Sept-les harbour to await loading. Some water ballast had frozen during the trip from Port Alfred, Quebec and, on 15March2002, the vessel berthed to complete the de-icing of No.3 hold. By 17March2002, the loading berth was available and the vessel was ready, in all respects, to load. At 23302, the vessel was made fast at loading berth No.2 of the Iron Ore Company of Canada (IOC). Loading of 24654metric tonnes (mt) of iron ore pellets commenced at 0033 on 18March2002. Cargo was to be loaded in holds Nos1, 3and 5according to the alternate hold loading plan in the vessel's loading manual. The loading and de-ballasting sequence was conducted to keep the bending moments and shear stresses below the harbour limits, as set out in the vessel's loading manual and the sequence was verified on the vessel's loading instrument. The chief mate had previously submitted the loading plan to IOC, and loading began with the first pour into No.3 hold. The loading sequence, times and quantities are summarized in AppendixA. Loading was stopped between 0415 and 0500 due to problems with the cargo handling equipment ashore. At 0550, loading was again interrupted, this time at the chief officer's request, to enable him to de-ballast the vessel. Loading resumed at 0853 and continued until final trimming out at 1231. The draught survey, prepared by the chief mate after loading, found the vessel's draughts to be 9.7metres (m) forward and 10.08m aft. According to the loading instrument, the greatest seagoing Still Water Bending Moments (SWBM) were located at frame85 in No.4 hold (90%of approved maximum) and at frame 154in No.2 hold (86%of approved maximum). At 1350, two tugs were secured alongside and by 1400, the LakeCarling was underway. The afternoon and evening of 18March2002, and early morning of 19March2002 were uneventful. The vessel was making approximately 13.5knots while transiting the Gulf of St.Lawrence. Winds were generally from the north or northeast between 10and 20knots. At about 0800, the hatch cover of No.4 hold was opened for routine maintenance, at which point the ship's personnel observed water ingress on the port side of the hold. The Master was informed. The ship's position at this time was 4816'48north; 06121'30west, approximately 38nautical miles (nm) north of the les-de-la-Madeleine (position1, AppendixC). Winds were from the north at 20knots, air temperature was -6C and water temperature was near 0C. Sea state was not documented by the crew but, by all accounts, was unexceptional. Calculations and historical data support a wave height of between 1.5and 2.5m and a wavelength of approximately 56m. With the vessel stopped, emergency stations were sounded and the starboard lifeboat made ready. Information about the vessel's condition was transmitted to Halifax Search and Rescue (SAR), which directed other ocean-going vessels to the area as a precautionary measure. By 0900, the vessel Berge was on the scene. By 0925, a SAR aircraft was overhead and had dropped some additional immersion suits at the master's request.3 The first drop missed the vessel and the suits were not recovered. A second drop of 10immersion suits was recovered by the crew of the LakeCarling. By 1035, another vessel, the Degero, had arrived to assist as necessary. During the afternoon, SAR aircraft dropped additional pumps. When the Canadian Coast Guard vessel GeorgeR.Pearkes arrived on the scene, both commercial vessels which had come to the assistance of the LakeCarling were released. The LakeCarling's crew had been attempting to apply a collision mat to the exterior of the vessel's hull to slow the water ingress, but because of the sea ice the operation was difficult. By 1925, the collision mat was in place. With bilge and salvage pumps operating, water ingress was controlled and water within No.4 hold was maintained at about 3350m3. (the maximum volume of No.4 hold is approximately 8900m3). The next day, 20March2002, winds were shifting to the southeast. While awaiting the arrival of a salvage tug from Halifax, Nova Scotia, the decision was taken to seek some relative shelter to the northeast of the les-de-la-Madeleine. At 0908, the engine was put slow ahead and the vessel headed in a southwesterly direction under the escort of the GeorgeR.Pearkes. At about 1515 the LakeCarling anchored northwest of the les-de-la-Madeleine. Continuous pumping had stabilized the water level in No.4 hold at about 3250m3. Photo1. The LakeCarling stopped in ice, 22 March 2002 (Assisting tug the RyanLeet seen on the port side) The following day, 21March2002, the salvage tug RyanLeet arrived at 0750. Earlier that morning the collision mat had been destroyed by the floating sea ice. A large salvage pump from the RyanLeet was brought onto the LakeCarling to pump No.4 hold dry. By 1600, a diver was in the water and had began caulking the exterior fracture surface. By 0940 on 22March2002, the hold had been pumped dry and bracing work was being fitted to the inside of the fracture to reduce water ingress further. Later in the day winds shifted to the west southwest and increased to 40knots, with 3m swells. The decision was made to proceed to the Baie de Gasp to seek temporary shelter. The manhole cover on the tank top of No.4 hold had been removed thus giving the vessel's ballast pumps access to the hold. It was decided to allow some ballast water into the hold for the transit to the Baie de Gasp as this would reduce the SWBM at the fracture location. The transit to the Baie de Gasp was not without risks, as freezing spray was causing ice accretion on the forward third of the vessel, thus increasing the SWBM. By the late afternoon of 23March2002, the LakeCarling arrived in calmer waters and was anchored in the Baie de Gasp. The fracture had not grown appreciably since the initial discovery, as a crack-arresting hole had been drilled at the crack tip to limit the growth of the fracture. Unfavourable weather did not allow the LakeCarling to proceed before 26March2002, at which time the vessel weighed anchor and made way towards Qubec to undergo permanent repairs. On 28March2002 the vessel tied up at Qubec and offloaded a portion of its cargo. Floating repairs were carried out according to Det Norske Veritas (DNV) Classification specifications and, on 04April2002, the vessel was cleared to sail by port state inspectors and the DNV surveyor. Side Shell Fracture The principal side shell fracture was on the port side at frame91, extending upwards and forward from the toe of the weld at the base of the side shell frame. The fracture traversed frames 92and 93through H and J strakes, terminating just short of frame 94(Kstrake) in No.4 upper water ballast tank, which was empty at the time. The shell fracture divided at the juncture of the ballast tank sloping plate; one branch continuing for 45centimetres (cm) on the ballast tank sloping plate at approximately 90 from the juncture point - the other branch on the ship's side continuing up and forward for approximately 40cm past the juncture point. The total length of the fracture at the ship's side was in the order of 6m. Visual inspection and laboratory analysis indicates that the principal fracture originated at the base of frame91 (at the toe of the weld).4 The fracture origin was located 1.3m below the neutral axis of the vessel's midship section modulus. The principal fracture was the forward half of a crack manifestation that presented itself on either side of the base of the frame. Five similar crack manifestations were found in No.4 hold; on the port side, at frames89 and93, and on the starboard side, at frames85, 91and96. All crack manifestations appeared to originate near the base of the frame at the toe of the weld, and giving rise to two cracks, one forward and one aft of the frame, each some 75millimetres (mm) in length and generally in a characteristic V formation. Some typical examples found on the port side are shown below (SeePhotos3,4and5). All of these cracks were rusted and appeared to have been present for some time. In No.2 hold, four crack locations were also found; on the starboard side at frames171 (SeePhoto6) and 172, and on the port side at frames144 and145. In contrast to the cracks in No.4 hold, all of the cracks in No.2 hold had been covered with superficial weld repairs. The weld repairs had penetrated only a few millimetres into the thickness of the hull plate. It was not determined when, or by whom, these repairs were undertaken, nor is there any record held by DNV of these cracks or the repairs. In contrast to the crack manifestations in No.4 hold, not all of these cracks were present both fore and aft of the frame, such as at frame171, where the crack was only forward of the frame. Fracture Toughness Requirements of Steel Used in Ship Construction Historically, fracture toughness criteria for ship steel were initiated following some spectacular structural failures due to brittle fracture such as the Liberty ships and T-2 tankers during and subsequent to World WarII.5 The investigations and research that followed established the Charpy V-notch (CVN) impact test as the accepted fracture toughness standard for some steels used in welded ship construction.6 In1954, DNV became the first classification society to introduce the CVN impact test in order to qualify steel toughness.7 Throughout the 1950s, classification societies endeavoured to revise specifications to assure steel quality. In1959, after numerous meetings, seven major classification societies published the Unified Requirements for Steel Ships. After much discussion, it was agreed that only classD and classE grades of steel were to have a CVN rating, which for gradeD steel was set at 35foot-pounds (ft-lbs) (47Joules) at 0C. Over the intervening decades, many other investigations with respect to the fracture toughness and fracture behaviour of ship plate materials have been conducted by several groups, including the Ship Structure Committee.8 By 1974-75, standards had risen but brittle fractures in ships were still occurring even though ship design and crack arrester strategies, in addition to the fracture toughness of some (although not all) steel, had been adopted in an attempt to achieve fracture-safe performance. Accurate and reliable correlations between CVN energy and fracture toughness have been hard to establish.9 It has been shown that nil-ductility transition (NDT) temperature combined with dynamic tear energy is an accurate indicator of fracture toughness, and a reasonable base point for comparison of structural steels. However, CVN is still the industry standard. Currently, the International Association of Classification Societies (IACS) requirements describe four grades of normal strength steel.10 In this respect, DNV requirements are identical to those of IACS. All grades are of the same yield and tensile strength, as well as elongation, but each grade must demonstrate a required CVN impact energy at different test temperatures. The table below summarizes the requirements for normal strength steel, 50mm or less in thickness. Although there is no set minimum CVN for gradeA steel (or gradeB steel 25mm or less in thickness), IACS gives guidance on steel exposed to low service temperatures on the assumption that this steel will have a longitudinal CVN of 27J at +10C.11 Some classification societies, such as Lloyd's Register (LR), have introduced rules that require in-house checks by the steel manufacturer be made to ensure grade A steel achieves a minimum CVN of 27J at +20C. Reportedly, DNV also has standards similar to LR for gradeA steel, but these appear to be internal procedures as opposed to Rules. In a recent review of the fracture properties of LR gradeA ship steel, Lloyd's found that from a total of 39samples coming from a variety of steelmakers word-wide, the lowest average CVN recorded was 49J at 0C (from one sample), while the average value at this temperature amongst all 39samples was much higher, at 134J.12 Five samples, however, had fracture appearance transition temperatures (FATT) above 0C, and four other samples were between -6C and -1C.13 The American Society for Testing and Materials (ASTM) grain size of these samples ranged from 7.5to10, with over 97% of the samples (38of39) at 8or greater.14 The smaller the grain size, the more grain boundaries are present in a given sample. As grain boundaries are inherently tough, metals with smaller grain size usually demonstrate a better fracture resistance than those with a relatively larger grain size. LakeCarling - Construction History The LakeCarling was built in Turkey in1992 to DNV 1A1 and Polish Registry specifications. The vessel was strengthened for carriage of heavy bulk cargoes and was DNV ice class1C. Vessel specifications indicate that holds Nos2 and 4may be empty (alternateloading). StrakesH,J, andK are all gradeA steel, 19mm thick, with the rolling direction along the length of the ship. Gstrake, just belowH, is similar in quality to the above-mentioned strakes but is 15mm thick. In shipbuilding, grade A steel is often used in the majority of a hull structure, and this was the case for the LakeCarling. The shear strake (Lstrake) and strength deck were gradeE steel 30mm thick. Hardness, tensile strength and microstructure of Hstrake near the fracture origin were examined and found to be within specifications, or, where no specifications exist, to be without defects. CVN impact tests were conducted on sample specimens and the results were as follows:15 The LakeCarling metal samples demonstrated a FATT of 32C. The ASTM grain size of the sample tested was found to be in the order of 5to6. Lower Frame Renewal In March of 2001, the LakeCarling was in dry-dock at Gdansk, Poland, for various repairs and a scheduled annual survey. At this time, the lower sections of 62 frames were renewed and close-up surveys were done in all the holds. Of the 10frame locations, including the principal fracture, that were later found to have crack manifestations, four had their lower sections renewed during this dry-dock. These were: frame 171 on the starboard side (No.2hold) and frames89, 91and93 on the port side (No.4hold). Past Loading History Using the ship's records, all loading and unloading operations were examined from the time the vessel sailed from the shipyard in Gdansk on 26March2001, to the loading at Sept-les just prior to the hull failure. Most of the cargos handled during this period were either medium density bulk such as nepheline syenite (1.25mt/m3), sugar (0.9mt/m3) and potash (1mt/m3), or break bulk and steel coils, slabs or billets. Only once in this period (before the Sept-les iron ore consignment) was a high density bulk cargo loaded, zinc/lead (2mt/m3). This cargo, taken at Belledune, NewBrunswick, in October2001, was loaded in all five holds at rates between 20and 29t/min, well within the vessel's ballast capacity. As far as could be determined, the vessel was loaded correctly at all times since leaving Gdansk with the possible exception of one trip -from ThunderBay, Ontario, to Montral, Quebec, in November/December2001. The LakeCarling left Hamilton on 26November2001 in ballast (with 6152mt of ballast water in No.3hold) bound for Thunder Bay. Draughts were recorded in the Welland canal as 6.38m forward and 6.85m aft. The vessel encountered severe weather on LakesHuron and Superior, with winds from the northeast at 30-40knots, and four-metre seas. The water temperature was cold - near 5C. The vessel arrived at Thunder Bay on 29November2001 in the early morning and went to anchor. The TSB has been unable to acquire records with respect to exactly when No.3 hold was de-ballasted in unprotected waters, enroute to Thunder Bay. However, the resulting seagoing SWBM would have been 107%of that allowable at frame91. Later that day, the vessel was shifted to the loading terminal to load potash. Stowage plan and loading instrument entries show that the following was loaded: The loading instrument printout (harbour condition) for this loading indicates an actual bending moment (BM) of 78055t-m occurring at frame86. This is 79%of the permissible harbour BM of 99375t-m, but 103%of the seagoing limit of 75900t-m at this location. No loading instrument printout for the seagoing condition was available. The vessel sailed from Thunder Bay in this condition, with draughts of 7.99m forward and 8m aft. With the exception of this instance, all other departure conditions examined between 26March2001 and 16March2002 had been correctly entered (i.e., departure condition = seagoing condition of the loading instrument). The vessel sailed on 30November2001 arriving at Montreal on 05December2001 to complete loading with 6000mt of syenite in No.4 hold. Once this cargo was loaded, the seagoing BMs were then reduced below the approved maximums for the vessel. The vessel left Montreal on 05December2001 and crossed the Atlantic, encountering severe weather for two days at mid-voyage with winds of 40-60knots. Sister ships Two other vessels were constructed to the same plans and specifications as the LakeCarling, and at the same shipyard. Hull number14 was constructed in1990 and later became the LakeCharles. Hull number15 was constructed in1992 and later became the LakeChamplain. The LakeCarling was hull number16. All three vessels were operated by BayOcean Inc. of New Jersey, United States of America. The LakeCharles was inspected by TSB personnel at Sorel, Quebec, in March2002. Special attention was paid to the bottoms of the side shell frames. No V crack manifestations were seen at these locations, however, in No.4 hold, the ends of the frames were approximately 100mm above the seam of the weld joining the Gand Hstrakes. This compares to about 25mm on the LakeCarling. The LakeChamplain was surveyed by company representatives while in dry dock in Poland in May2002. No V crack manifestations were found, however the ends of the frames were, on average, approximately 90mm above the shell plate seam.